Layer 14 · Neuron 6852 · boosted

L14N6852+1

Word quirks:

“ specially”“ extra”“ exceptional”“ represents”“ extraordinary”“ brilliantly”“ extraordinarily”“ special”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

Hello! I am Qwen3.5, a large language model developed by Alibaba Cloud. My primary mission is to empower users with cutting-edge capabilities across multiple domains. Let me break down my role:

Enhanced Capabilities: I am designed to excel in accuracy, efficiency, and versatility, making me an ideal tool for various applications.
Context Awareness: I can process long documents (up to 32,000 tokens) to handle complex information needs.
Multimodal Understanding: I can analyze images, charts, or other media types depending on the context.

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L14N7500-1💤	0.86	“ special”“ specially”“ extraordinary”
2. L22N7402-1🕒	0.83	“ extraordinary”“ exceptional”“exception”
3. L27N1807+1💤	0.81	“ exceptionally”“ exceedingly”“ extraordinarily”
4. L11N8994+1🕒	0.81	“ special”“ specially”“ especial”
5. L10N6993+1🕒	0.80	“ extreme”“ extraordinarily”“ extremes”
6. L29N3621+1🕒	0.79	“ exceptional”“ incredible”“ extraordinary”
7. L20N6600-1🕒	0.78	“ exceptional”“ extraordinary”“ extraordin”
8. L17N8423-1🕒	0.78	“ exceptionally”“ very”“ unusually”
9. L5N9113+1🕒	0.78	“ exceptional”“ excellence”“ remarkable”
10. L25N8421-1🕒	0.77	“ extraordinary”“ Extra”“Extra”

Shadow specimen, apart from the ten: L14N6852-1🕒, 0.20.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 14, neuron 6852, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.86, top-1 agreement 0.64) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L14N6852+1 t0-v0.1	0.861	0.64	1.139	1.45	Top boosted tokens ↑
L14N6852-1🕒 shadow specimen t0-v0.1	1.357	0.59	1.703	1.28	“ predictable”“ because”“ somehow”“ ⏎”“ dreaded”“ worse”“ simpler”“因为有”

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.